2 BOTANICAL MAGAZINE. vol.xvn, 



specimens suspected to be fructified, I searched carefully after those cells, but 

 iu vain. E. M. Holmes^^ claims to have found an abundance of antheridia 

 in his specimens, but that this is not the case is manifest from his illustra- 

 tions (PI. VIII. Fig. 4 a and c). He figures a branch bearing anthenidia, 

 each of which is so large as it is visible to naked eyes in the natural size. 

 No one shall believe that there is found such a large antheridium in plants 

 among Rhodomelaceae. If it were really antheridium, as Holmes claims, his 

 plant would have no doubt been a new genus, there being no need to 

 " stand provisionally until specimens with cystocarps and tetraspores are 

 obtainable/^ He does not give further details with respect to his antheridia 

 except their obovate form, in spite of his having noticed them to " difier 

 entirely in character from those of OhondriaJ' The bodies, which that 

 author supposes to be antheridia, are either the beginnings of sterile and 

 normal branches or those which I describe as knobs below. They are oblong 

 and appear, especially when young, to careless eyes like antheridia in 

 Rhodomelaceae. 



This alga has many minute knob-like ramelli which are produced on 

 the apices of ramuli. Those knobs are elongato-obovate or ellipsoidal in 

 shape and are formed 5-7 or more on one and the same apex. Those 

 knobs, when fully grown, become so much swollen into somewhat firm bodies, 

 that they do not flatten during the process of drying, and their size is 

 generally 1 mm. in length and 0.5 mm. in breadth. They are found to 

 be formed in summer and ripen during winter. In the summer of 1901, 

 when I was studying the structure of this alga at Kadzusa for determining 

 the systematic position for it, a doubt occured in my mind whether those 

 knobs, in which we had hitherto been endeavoring to find out either of its 

 sexual cells, would not themselves be substituted for the propagative organs. 

 Struck with this supposition, I examined many well-formed knobs and found 

 that the cells of the intermediate layer, that is the layer interposed between 

 the epidermis and the central axis is richly filled up with globular starch 

 grains. Even some of the epidermal cells were furnished with them. The 

 knob is constructed on the same structure as any other portion being fur- 

 nished with a central axis, which ends in an apical cell, and five pericentral 

 cells. In outside of the latter, we find a thick cortical layer of roundish 

 cells which are formed by branching from them, being covered finally by a 

 single layer of cortical cells. Thus, the structure of knobs differs from the 



]). ?Chondria crassicaulis Harv. in Holmes New Mar. Algae from Japan. (Journ. Linn. 

 Soc— Bot. Vol. XXXr. ). p. 256 PI. VIII Fig.:4 a, 6, c. 



